rework external process creation action validation. Publications now override the creation action property making it "generated". Checks for ancestor entity type now occurs within dedicated neo4j query

Author: DerekFurstPitt

src/schema/provenance_schema.yaml

@@ -654,6 +654,15 @@ ENTITIES:
         indexed: true
         description: "The title of the publication."
         required_on_create: true # Only required for create via POST, not update via PUT
+      creation_action:
+        type: string
+        transient: true
+        generated: true
+        immutable: true
+        indexed: true
+        on_read_trigger: get_creation_action_activity
+        on_index_trigger: get_creation_action_activity
+        description: "The activity that was performed."
       dataset_type:
         before_create_trigger: set_publication_dataset_type
         before_property_create_validators:

src/schema/schema_neo4j_queries.py

@@ -112,6 +112,37 @@ def get_entity(neo4j_driver, uuid):
     return result
 
 
+
+"""
+Get the uuids for each entity in a list that doesn't belong to a certain entity type. Uuids are ordered by type
+
+Parameters
+----------
+neo4j_driver : neo4j.Driver object
+    The neo4j database connection pool
+direct_ancestor_uuids : list
+    List of the uuids to be filtered
+
+Returns
+-------
+dict
+    A dictionary of entity uuids that don't pass the filter, grouped by entity_type
+"""
+def filter_ancestors_by_type(neo4j_driver, direct_ancestor_uuids, entity_type):
+    uuids_str = ", ".join([f'"{uuid}"' for uuid in direct_ancestor_uuids])
+    query = (f"MATCH (e:Entity) "
+             f"WHERE e.uuid in [{uuids_str}] AND toLower(e.entity_type) <> '{entity_type.lower()}' "
+             f"RETURN e.entity_type AS entity_type, collect(e.uuid) AS uuids")
+    logger.info("======filter_ancestors_by_type======")
+    logger.info(query)
+
+    with neo4j_driver.session() as session:
+        records = session.run(query).data()
+          
+    return records if records else None
+   
+
+
 """
 Get all children by uuid
 

src/schema/schema_validators.py

@@ -589,12 +589,11 @@ def validate_creation_action(property_key, normalized_entity_type, request, exis
         if normalized_entity_type.lower() != "dataset":
             raise ValueError(f"Only datasets are allowed to have creation_action of 'external_process'. Provided entity is a {normalized_entity_type}")
         direct_ancestor_uuids = new_data_dict.get('direct_ancestor_uuids')
-        for ancestor_uuid in direct_ancestor_uuids:
-            ancestor = schema_neo4j_queries.get_entity(schema_manager.get_neo4j_driver_instance(), ancestor_uuid)
-            ancestor_type = ancestor['entity_type']
-            if ancestor_type.lower() != "dataset":
-                raise ValueError(f"If creation_action is 'external process', all direct ancesotrs must be of type 'dataset'. Ancestor with uuid {ancestor_uuid} has entity type of '{ancestor_type}'")
-           
+        entity_types_dict = schema_neo4j_queries.filter_ancestors_by_type(schema_manager.get_neo4j_driver_instance(), direct_ancestor_uuids, "dataset")
+        if entity_types_dict:
+            raise ValueError(f"If 'creation_action' field is given, all ancestor uuids must belong to datasets. The following entities belong to non-dataset entities \
+                             {entity_types_dict}")
+        
 
 """
 Validate the provided value of the activity creation action before updating direct ancestors. Certain values prohibited